On a hackathon I have been to recently, one team was pitching an idea for a mobile App that helps you chose your future kids name together with your partner. The technical solution they were proposing was by having each partner rate names in a binary yes-no-fashion and using recommendation systems to find a name both like equally much.
Aside from the obvious short-comings of binary-yes-no-approach for this solution, there is an even higher, conceptional problem with it. Yes, sure, it is a problem for those in that situation. And it is a hard and probably very nasty problem with high responsibility - after all that person will carry this decision for the rest of their life - but it is also a highly personal one. One that you never want to solve technically.
Just picture the following dialogue. The seven year old kid runs to their parents after their friend told them the amazing story of how their parents chose their name and very excited asks: “Mommy, mommy, my name - where does it come from? Why did you pick this name?” and Mommy replies “Oh, there is an app for that.” - that poor kid.
But it demonstrates a very important factor you should always take into account when thinking of the solution of a certain problem: is this is a problem that allows a technical solution the first place? And many highly emotional and super personal things do not. That is why dating sites suck or music recommendation is a tough topic: it is highly personal and emotional and while we allow machines to help us with the process we could never allow them to decide for us.
We will never like their decision as much as if they gave us the choice at the end, even if they are more right about us than we are. We would simply feel miserable about that sound this machine found for us in comparison to the song we found.